Biochemistry 1992, 31, 9040-9044
9040
Determinants of High-Affinity DNA Binding by the Glucocorticoid Receptor: Evaluation of Receptor Domains outside the DNA-Binding Domain? Karin Dahlman-Wright,' Anthony P. H. Wright, and Jan-Ake Gustafsson Centre for Biotechnology and Department of Medical Nutrition, Karolinska Institute, NO W M , Huddinge University Hospital, S- I41 57 Huddinge. Sweden Received March 2, 1992; Revised Manuscript Received June 18, 1992
In this study, we have investigated the influence of regions outside the DNA-binding domain of the human glucocorticoid receptor on high-affinity DNA binding. We find that the DNA-binding domain shows a 10-fold lower affinity for a palindromic DNA-binding site than the intact receptor. The N-terminal part of the receptor protein does not influence its DNA-binding affinity, while the C-terminal steroid-binding domain increases the DNA-binding affinity of the receptor molecule. It has previously been shown that both the intact glucocorticoid receptor and the glucocorticoid receptor DNA-binding domain bind to a palindromic glucocorticoid response element on DNA as dimers. It is likely that differences in DNA-binding affinity observed result from protein-protein interactions outside the DNA-binding domain betwen receptor monomers, as has been shown for the estrogen receptor. We have previously identified a segment involved in protein-protein interactions between DNA-binding domains of glucocorticoid receptors. This, in combination with results presented in this study, suggests that there are at least two sites of contact between receptor monomers bound to DNA. We suggest that the interaction between the DNA-binding domains may act primarily to restrict DNA binding to binding sites with appropriate half-site spacing and that additional stability of the receptor dimer is provided by the interactions between the steroid-binding domains. ABSTRACT:
The glucocorticoid receptor (GR)' belongs to a family of nuclear receptors which activate transcription in response to binding of their respective ligands. Transcriptional activation is accomplished by DNA-bound receptors. DNA-bindingsites for the GR are called glucocorticoid response elements (GREs) and are characterized by palindromic half-sites which are separated by a three amino acid spacer of variable sequence [forreviews,seeGustafssonetal. (1987),Evans (1988),Beato (1989), and OMalley (1990)l. The palindromic nature of their response elements suggests interaction with receptor dimers, and this has been demonstrated for the GR (Tsai et al., 1988; Wrange et al., 1989; Chalepakis et al., 1990), the estrogen receptor (ER) (Kumar & Chambon, 1988; Fawell et al., 1990), and the thyroid hormone receptor (TR) (Brent et al., 1991; Selmi & Samuels, 199l;Lazaret al., 1991). TheGR(Wrangeet al., 1989), the ER (Kumar & Chambon, 1988; Fawell et al., 1990), and the progesterone receptor (PR) (Rodriguez et al., 1990; DeMarzoet al., 1991)alsoshow somedimerizationin theabsence of DNA. The ER is the best characterized steroid receptor with respect to the segments involved in receptor dimerization. Using ER expressed in yeast cells, it was shown that two segments of the ER are involved in dimerization: one in the steroid-binding domain and one in the DNA-binding domain (Kumar & Chambon, 1988). However, the major dimer?This work was supported by grants From the Swedish Medical Research Council (2819) and the Swedish Natural Science Research Council (K-KU 9756-301). Salary support was provided by the Swedish Medical Research Council to K.D.-W. (student fellowship). * To whom correspondence should be addressed. Telephone: +46 8 6089158. Telefax: +46 8 7745538. Abbreviations: GR, glucocorticoidreceptor;ER, estrogen receptor; TR, thyroid hormone receptor; PR, progesterone receptor;DBD, DNAbinding domain; GRE, glucocorticoid response element; ERE, estrogen response element; DTT, dithiothreitol; TAT, tyrosine aminotransferase.
0006-2960/92/043 1-9040$03.00/0
ization activity was shown to reside in the steroid-binding domain (Kumar & Chambon, 1988; Fawell et al., 1990). A more detailed characterization has revealed the importance of amino acids at the C-terminal end of the steroid-binding domain for ER dimerization (Fawell et al., 1990). The corresponding segment has also been implicated in heterodimerization between retinoic acid and thyroid hormone receptors (Glass et al., 1989). For the PR, mutagenesis studies have indicated a dimerization activity within the ligand-binding domain (Guiochon-Mantel et al., 1989). Finally, it has been suggested that the N-terminal domain of the GR mediates protein-protein interactions between molecules constituting receptor dimers (Eriksson & Wrange, 1990). Two molecules of the GR DNA-binding domain (DBD) bind to a GRE cooperativelydue to protein-protein interactions between bound DBDs (Dahlman-Wright et al., 1990). Mutation of a five amino acid segment close to the second zinc ion has been shown to abolish this activity (Dahlman-Wright et al., 1991). Transient transfection studies using mutated receptor derivatives have shown that this five amino acid segment functions to restrict the transactivation capacity of the glucocorticoid receptor to response elements where the distance between the half-sites is three base pairs (Umesono & Evans, 1989). Thus, protein-protein interactions between DBDs within GR dimers are involved in the specificity of binding site selection. This study was initiated to determine whether the interactions between GR DBDs, in addition to their role in binding site selection, are the sole determinants for GR dimerization. For the ER, it has been shown that receptor dimerization is required for high-affinity DNA binding (Kumar & Chambon, 1988; Fawell et al., 1990). In this study, we have measured the DNA-binding affinity of various GR derivatives to evaluate the role of DBD in relation to other receptor domains in receptor dimerization. 0 1992 American Chemical Society
Biochemistry, Vol. 31, No. 37, 1992 9041
Glucocorticoid Receptor DNA Binding
EXPERIMENTAL PROCEDURES Plasmid Constructions. The yeast expression vector, pKV50, used for expression of GR derivatives has been described (Wright et al., 1990). Using this vector, proteins are expressed from the PGK promoter, the upstream activation sequences of which have been replaced with those from the GAL1,10 promoter. In this expression system, expression is repressed during growth on glucose-containing media and induced on galactose-containing media. The plasmids pKVNX, pKVXE, pKVAX, and pKVNXAT1 have been described previously (Wright & Gustafsson, 1991)and express the following receptor derivatives, respectively: GR (intact receptor, residues 1-777); GRDBD (residues 37&500); GRNtermA (residues 415-777); and GR71A (residues 77-262 deleted). The plasmid pKVNE encodes the GRCtermA protein (residues 1-500) followed by Asp-His-Gly at the Cterminus. Plasmids were transformed into the Saccharomycescerevisiae strain W303-1A (MATa,ade2-1, canl-100, his3-1 I ,15, leu2-3,112, trpl-1, ura3-1) according to Beggs (1978). Growth of Yeast Strains and Preparation of Extracts. Yeast strains were grown to late log phase in SD medium (Sherman et al., 1986) without leucine at 30 OC, harvested by centrifugation, and washed in the same medium without added carbon source. The cells were harvested and resuspended in SGE medium (0.67% yeast nitrogen base without amino acids, 3% glycerol, and 1% ethanol) without leucine, at a density of A600 = 0.5. After 24-h growth at 30 OC, during which glucose repression was relieved, the cells were diluted to a density of A600 = 0.5 in SGE medium without leucine, and galactose was added to a final concentration of 2% to induce expression of the recombinant proteins. Cells were harvested after 8-h growth at 30 "C. Yeast extracts were prepared essentially as described (Wright et al., 1990) except that the extraction buffer was 200 mM Tris-HCI, pH 8.0,400 mM (NH&S04,10 mM MgC12,l mM EDTA, 10%glycerol, 2 mM DTT, and 1 mM PMSF. Extracts were dialyzed against 10 mM Hepes, pH 8.0,10% glycerol, 2.5 mM MgC12,50 mM KCI, 0.1 mM EDTA, and 1 mM DTT. Gel Mobility Shift Assays. For DNA-binding assays, 5 pL of extract was incubated in 10 mM Hepes, pH 8.0, 10% glycerol, 2.5 mM MgC12,50 mM KC1,O. 1 mM EDTA, 1 mM DTT, 1 pg of poly(d1-dC)*poly(dI-dC), and 100 ng of a 41 base pair palindromic estrogen response element (ERE1) described previously (Zilliacus et al., 1991) with various labeled DNA fragments in a total volume of 10 pL. DNA concentrations were determined spectrophotometrically, assuming A260 = 1 correspondsto 20 pg/mL oligonucleotide. The GRE in Figure 1 is derived from the tyrosine aminotransferase (TAT) gene (Jantzen et al., 1987) and is 41 base pairs in length. Oligonucleotideswith inserted or deleted base pairs were correspondingly longer and shorter (Dahlman-Wright et al., 1990). The palindromic GRE studied in Figures 2-5 is a 41 base pair palindromic GRE (GRE1) described previously (Zilliacus et al., 1991). In Figures 1 and 2B, 0.2 ng of labeled oligonucleotidewas added to each incubation. In Figures 3-5, the amount of oligonucleotide added is indicated in the figures. High concentrations of DNA were achieved by diluting the labeled oligonucleotidewith nonlabeled oligonucleotide. Protein-DNA complexes were separated from protein-free DNA by nondenaturingelectrophoresis (Fried & Crothers, 1981) in 5% polyacrylamide gels. The gels were run at 4 OC in 50 mM Tris-boric acid, pH 8.0, 1 mM EDTA, and 0.1 % Triton X- 100 at a constant voltage of 200 V. Following electrophoresis, gels were dried and au-
GRE
A
h
GREM4
GREM5
GREM6
GREM12
h
h
h
h
am--
---
C-, a
I GREM4
(-1GAE)
GREM5
(-[GAGA=)
GREMG
(K~GGACC
FIGURE1: GR dimerization correlates with high-affinity DNA binding. (A) Binding of intact GR to the TAT GRE or variants of this GRE was assayed by band shift analysis. The variant GREs contain 2 (GREM4), 4 (GREMS), or 13 (GREM6) base pairs between the half-sites or have an altered relative orientation of the half-sites (GREM 12). F and C show the positions of the free probe and the GR-DNA complex, respectively. Extracts are diluted 2fold serially from left to right. (B) Sequence of the variant GREs studied in (A). toradiographed. Data were quantified by scintillation counting of bands excised from gels. In some cases, the bands were also quantified by densitometryof autoradiographswhich gave very similar results. The amount of DNA bound was calculated from the specific activity of the probe. Scatchard analysis was used to calculate the dissociation constants for the different proteins. The values measured were reproducible, with individual values varying from the mean of two or more experiments by about 5%, except for the GRNtermA protein where thevariation was 18% from the mean of two experiments. The R values show the reliability of the regression lines used in the Scatchard plots. The Scatchard plots for the GRDBD and GRCtermA proteins are nonlinear at low DNA concentrations, due to cooperativeDNA binding, and, consequently, these points were excluded from the analysis.
RESULTS G R Dimerization Is Required f o r High-Affinity DNA Binding. To evaluate the role of GR dimerization for highaffinity DNA binding, binding of intact GR expressed in yeast cells to a GRE from the TAT gene and variants of this GRE, where the distance between the two half-sites or their relative orientation is changed, was studied using a gel retardation assay (Figure IA). The sequences of the TAT GRE and the variant GREs are depicted in Figure 1B. Binding of the GR, expressed in yeast cells, resulted in the complex marked C in Figure 1A which is specific to the GR since extracts expressing
9042
Biochemistry, Vol. 31, No. 37, 1992
A
DBD
11 r
GR
-_1_1--
.
R
Dahlman-Wright et al.
SBD
1 - y'--
GRDBD GRCtermA
-
I -
-
GRNtermA
B
i
2
3
-"-----T
4
5
o 1.5
e
+ 4:
E 1.0 \ 'D c
;0.5
n
9 0 -.nn
1 .o
0.5
0.0
DNA bound (nM)
. . - - -. . - - - .. . . .
-.
15.0.
;c
A
10.0 .
A
A
5.0 .
: A --..
- * - . .
5
10
. - - - . ' - - . .- -.-
15
20
25
30
DNA (ng)
FIGURE2: Sequence-specific DNA binding by GR derivatives expressedin yeast. (A) GR derivativesexpressed in theyeast extracts utilized in this study. The steroid-binding domain (SBD), the DNAbinding domain (DBD), and the transactivation domains (71 and 72) are indicated. (B) Complexes formed by GR derivatives expressed in yeast using a palindromic GRE (AGAACATGATGTTCT) in a gel mobility shift assay. Extracts contained the expressed proteins as follows: GR (lane l), GRDBD (lane 2), GRCtermA (lane 3), GRNtermA (lanel),and GR71A (lane 5 ) . Fand C show thepositions of the free probe and the protein-DNA complexes,respectively. Note that the lanes shown should not be compared quantitatively since they are taken from DNA titration experiments and contain different amounts of protein and/or DNA.
smaller GR derivatives do not give this complex (Wright & Gustafsson, 1991). Changing the distance between the halfsites constituting a GRE, or changingtheir relative orientation, changes the relative position of receptor molecules on the DNA, potentially interfering with interactions between receptor monomers. Figure 1A shows that the affinity of GR for DNAbinding sites where the distance between the half-sites or their relative orientation has been changed is much reduced. This is consistent with the binding behavior previously shown for the GR DBD (Dahlman-Wright et al., 1990) and supports theview that a dimerized receptor structure, that is compatible with specific interaction with both half-sites of the GRE, is required for high-affinity DNA binding of the GR contained in crude yeast extracts. A similar correlation between halfsite spacing and DNA-binding affinity has previously been demonstrated for GR purified from rat liver (Chalepakis et al., 1990). DNA-BindingAfJinitiesof the Intact GR and the GR DBD. To evaluate the role of the protein-protein contact within the DBD for receptor dimerization and high-affinity DNA binding, the DNA-binding affinities of the intact GR and the GR DBD were compared. The GR and GRDBD proteins expressed in yeast are depicted in Figure 2A. Western blot analysis showed that these proteins, as well as the other GR derivatives used in the study, appeared to be intact in the
3.0
- -
. . . .- . . . . . . . .
:"\ 2.51
D
Kd=5.7 nM
1 .o
0.050
5
10
15
20
DNA bound (nM)
FIGURE 3: DNA-binding affinity of the intact GR and the GR DBD. Scatchard analysis of gel shift assays with yeast extracts containing the GR (B) and the GRDBD (D) proteins. The DNA fragment was a palindromicGRE with sequenceAGAACATGATGTTCT. Shown are the saturation curves for GR and GRDBD (A and C, respectively) and the Scatchard plots for GR and GRDBD (B and D, respectively). The R values for the regression analysis of the Scatchard data were 0.94 and 0.98, respectively.
extracts (data not shown). The DNA-binding activity of the expressed proteins was analyzed using a gel retardation assay. As can be seen in Figure 2B, incubation of extracts expressing GR (lane 1) or GRDBD (lane 2) with a palindromic GRE results in formation of a protein-DNA complex. The migration of the complexes varies according to the size of the receptor derivative in the extracts, and therefore they reflect specific interactions with the GRE. The palindromic GRE used for the affinity determinations was chosen because it shows a lower level of nonspecific binding compared to the TAT GRE (compare Figure 2B and Figure 1A). The DNA-binding affinitiesof the GR and GRDBD proteins were compared by Scatchard analysis of gel retardation experiments using the palindromic GRE (Figure 3). The DNA-binding affinity of the intact GR is about 10-fold higher than that of the isolated DBD, Kd = 0.5 nM versus Kd = 5.7 nM, respectively. This shows that sequencesoutside the DBD are required for high-affinity DNA binding by the GR. These sequences probably mediate protein-protein contacts between receptor monomers that stabilize the protein-DNA complex.
Biochemistry, Vol. 31, No. 37, 1992 9043
Glucocorticoid Receptor DNA Binding
5
: v
5.0 '
O
.
O
F
l
n
n
z
0.3 0.2
2 0.1 0.0 0
5
10
15
20
25
A
0.0 0.0
30
0.5
DNA (no)
5
2.0
Kd-0.95 nM
1.0
E
0.0 0
a i0 0.0 n
0.5 0
1.5
1.0 DNA ( n d
5
10
15
FIGURE 4: DNA-binding affinity of the C-terminally deleted GR. Scatchard analysis of gel shift assays with a yeast extract containing GRCtermA (B). The DNA fragment was a palindromic GRE with sequence AGAACATGATGTTCT. Shown are the saturationcurve (A) and the Scatchard plot (B). The R value for the regression analysis of the Scatchard data was 0.99.
-?
v
I
A
1 .o
Table I: DNA-Binding Affinity of Intact and Truncated OR Proteins GR derivative Kd (nM) GR derivative Kd (nM) GR 0.5 GRNtermA 0.95 GRDBD 5.7 GRT~A 1.4 GRCtermA 5.5
0.0 0.0
1.0
A , 2.0
,
,
3.0
4.0
j
5.0
DNA (no)
ID
-
3.0
Having established the importance of sequences outside the DBD for high-affinity DNA binding, we investigated whether the N-terminal and/or the C-terminal part of the receptor protein contributed to the difference in affinity between the intact GR and the GR DBD. For the ER, the C-terminal steroid-binding part of the receptor protein has been shown to contribute to receptor dimerization and highaffinity DNA binding (Kumar & Chambon, 1988; Fawell et al., 1990). However, for the GR, it has been suggested that protein-protein interactions involving the N-terminal part of the receptor play a role in the interaction of the dimeric receptor with DNA (Eriksson & Wrange, 1990). Role of the G R N- Terminus in High-Affinity DNA Binding. The N-terminal part of the GR (GRCtermA, Figure 2A) was expressed in yeast. Incubation of the extracts with a labeled GRE fragment led to the formation of a protein-DNA complex (Figure 2B, lane 3). This complex represents a specific DNAprotein complex since it was not seen in extracts expressing other GR derivatives (Figure 2B). Quantification of gel retardation experimentsfollowed by Scatchard analysis shows that the DNA-binding affinity of the GRCtermA protein is very similar to that of the GRDBD protein, Kd = 5.5 nM versus Kd = 5.7 nM, respectively (Figures 3 and 4). This shows that sequences N-terminal of the GR DBD do not contain an independent activity that increases the DNAbinding affinity of the receptor protein. Role of the G R C-Terminus in High-Affinity DNA Binding. The lack of dimerization activity in the N-terminal domain suggests that the C-terminal ligand-binding domain of the GR contributes to receptor dimerization, as has been shown for the ER (Kumar & Chambon, 1988; Fawell et al., 1990). To test this directly, receptor derivatives from which the whole N-terminus (GRNtermA) or part of the N-terminus (GR71A) were deleted, shown in Figure 2A, were expressed and the
IC
'IAA, 3.0
1 .o
0.5 DNA bound (nM)
0.0
DNA bound (nM)
0.0 0.0
1.0
2.0
3.0
4.0
5.0
DNA bound (nM)
FIGURE5: DNA-binding affinity of N-terminally deleted GR derivatives. Scatchard analysis of gel shift assays with yeast extracts containing the proteins GRNtermA (B) and GRrlA (D). The DNA fragment was a palindromic GRE with sequence AGAACATGATGTTCT. Shown are the saturation curves for GRNtermA and G R T ~ A(A and C, respectively) and the Scatchard plots for GRNtermA and GRTlA (B and D, respectively). The R values for the regression analysis of the Scatchard data were 0.96 and 0.97, respectively.
extracts used to produce the specific complexesshown in Figure 2B (lanes 4 and 5 ) . These proteins bound to the GRE with dissociation constants of 0.95 and 1.4 nM, respectively, indicatingthat the C-terminus does contribute to high-affinity DNA binding by the GR. The two proteins gave very similar results in spite of the low level of active protein in the GRNtermA extracts. This probably results from aggregation of the GRNtermA protein when it is overexpressed (Wright et al., 1990) and causes an increase in nonspecific binding to the GRE (Figure 2B), thus making quantitation more difficult. DISCUSSION
In Figure 1A, we show that high-affinity DNA binding by the GR is practically abolished when the distance between, or the relative orientation of, the two half-sites constituting the GRE is altered. This probably reflects an influence on the ability of the receptor to form DNA-bound dimers as has previously been suggested (Chalepakis et al., 1990). We have used a gel retardation assay, followed by Scatchard analysis, to evaluate the role of different parts of the GR
9044
Biochemistry, Vol. 31, No. 37, 1992
for maximal DNA-binding affinity, as summarized in Table I. The results show that regions outside of the DBD are required for maximal DNA-binding affinity by the GR (Figure 3). It is likely that this increased DNA-binding affinity is the result of protein-protein interactions between receptor monomers bound to the GRE. This has previously been demonstrated for the ER (Kumar & Chambon, 1988; Fawell et al., 1990). However, we cannot exclude that regions outside the DBD mediate direct contactswith DNAor that they indirectly influence DNA-binding affinity by, for example, causing structural changes within the DBD. The N-terminal region of the GR has previously been implicated in receptor dimerization (Eriksson & Wrange, 1990). However, this domain does not appear to increase the DNA-binding affinity of the receptor and thus would not be expected to mediate contacts between receptor monomers (Figure 4), at least not in the absence of the steroid-binding domain. Addition of the C-terminal steroid-binding domain increases the DNA-binding affinity of the DBD (Figure 5). This suggests that this part of the receptor molecule contributes protein-protein interactions as has previously been shown for the ER (Kumar & Chambon, 1988; Fawell et al., 1990). The DNA-binding affinity of the GRNtermA and GR71A proteins is not quite as high as that of the intact GR. Thus, the N-terminus seems to be required to achieve maximal DNAbinding affinity. This could reflect an auxiliary role of the N-terminus that enhances the activity in the C-terminus and would be consistent with previous observations that the N-terminus does play a role in receptor dimerization (Eriksson & Wrange, 1990). Alternatively, since the contribution of the N-terminus is very small, it might be due to a general artifact associated with theuse of abnormal truncated proteins. At present, we cannot distinguish between these possibilities. However, we can conclude that the C-terminus of the GR makes a major contribution to the receptor’s DNA-binding affinity and that the N-terminus cannot contribute independently, although it might perform an auxiliary function in the intact GR. In vitro mutagenesis has revealed the importance of amino acids at the C-terminal end of the steroid-binding domain for ER dimerization. A striking feature of this sequence is a heptad repeat of hydrophobic residues with additional hydrophobic residues at intermediate positions (Fawell et al., 1990). The segment including these amino acids, particularly the repeating pattern of hydrophobic residues, is conserved within the nuclear receptor family, and it will be interesting to see if the same amino acids are involved in high-affinity DNA binding by the GR.
Dahlman-Wright et al.
REFERENCES Beato, M. (1989) Cell 56, 335-344. Beggs, J. (1978) Nature 275, 104-108. Brent, G. A., Williams, G. R., Harney, J. W., Forman, B. M., Samuels, H. H., Moore, D. D., & Larsen P. R. (1991) Mol. Endocrinol. 5, 542-548. Chalepakis,G., Schauer, M., Cao, X.,& Beato, M. (1990) DNA Cell Biol. 9, 355-368. Dahlman-Wright, K., Siltala-Roos, H., Carlstedt-Duke, J., & Gustafsson, J.-A. (1990) J. Biol. Chem. 265, 14030-14035. Dahlman-Wright,K., Wright, A., Gustafsson,J.-A., & CarlstedtDuke, J. (1991) J. Biol. Chem. 266, 3107-3112. DeMarzo, A. M., Beck, C. A., Onate, S.A., & Edwards, D. P. (1991) Proc. Natl. Acad. Sci. U.S.A. 88, 72-76. Eriksson, P., & Wrange, 0. (1990) J. Biol. Chem. 265, 35353542.
Evans, R. M. (1988) Science 244, 889-895. Fawell, S. E., Lees, J. A., White, R., & Parker, M. (1990) Cell 60, 953-962.
Fried, M., & Crothers, D. M. (198 1) Nucleic Acids Res. 8,65056525.
Glass, C. K., Lipkin, S. M., Devary, 0.V., & Rosenfeld, M. G. (1989) Cell 59, 697-708. Guiochon-Mantel, A., Loosfelt, H., Lescop, P., Sar, S., Atger, M., Perrot-Applanat,M., & Milgrom,E. (1989) Ce1157,11471 154.
Gustafsson, J.-A., Carlstedt-Duke, J., Poellinger, L., Okret, S., Wikstrbm,A X . , Brbnnegiird, M., Gillner, M., Dong, Y.,Fuxe, K., Cintra, A,, Hlrfstrand, A., & Agnati, L. (1987) Endocr. Rev. 8, 185-234. Hollenberg, S. M., Weinberger, C., Ong, E. S., Cerelli, G., Oro, A., Lebo, R., Thompson, E. B., Rosenfeld, M. G., & Evans (1985) Nature 318, 635-641. Jantzen, H.-M., Strlhle, U., Glass, B., Stewart, F., Schmid, W., Boshart, M., Miksicek, R., & Schiitz, G. (1987) Cell 49,2938.
Kumar, V., & Chambon, P . (1988) Cell 55, 145-156. Lazar, M. A., Berrodin, T. J., & Harding, H. P. (1991) Mol. Cell. Biol. 11, 5005-5015. Ohara-Nemoto, Y.,Strbmstedt, P.-E., Dahlman-Wright, K., Nemoto, T., Gustafsson, J.-A., & Carlstedt-Duke, J. (1990) J. Steroid Biochem. Mol. Biol. 37, 481-490. O’Malley, B. W. (1990) Mol. Endocrinol. 4, 363-369. Rodriguez, R., Weigel, N., O’Malley, B., & Schrader, W. T. (1990) Mol. Endocrinol. 4, 1782-1790. Selmi, S., & Samuels, H. H. (1991) J. Biol. Chem. 266,1158911593.
Sherman, F., Fink, G. R., & Hicks, J. B. (1986) Laboratory Course Manualfor Methods in Yeast Genetics, 2nd ed., Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. Tsai,S., Carlstedt-Duke,J., Weigel, N., Dahlman, K., Gustafsson, J.-A., Tsai, M.-J., & O’Malley, B. (1988) Cell 55, 361-369. Umesono,-K., & Evans, R. M. (1989) Cell 57, 1139-1146. Wrange, O., Eriksson, P., & Perlmann, T. (1989) J. Biol. Chem. 255, 5253-5259.
ACKNOWLEDGMENT We thank the members of the Steroid Receptor Unit at the Centre for Biotechnology for useful discussions and advice, Helmi Siltala for excellent technical assistance, and Delta Biotechnology Limited for the pKV5O yeast expressionvector.
Wright, A. P. H., & Gustafsson, J.-A. (1991) Proc. Natl. Acad. Sci. U.S.A. 88, 2883-2887. Wright, A. P. H., Carlstedt-Duke, J., & Gustafsson,J.-A. (1 990) J. Biol. Chem. 265, 14763-14769. Zilliacus, J., Dahlman-Wright, K., Wright, A,, Gustafsson, J.A., & Carlstedt-Duke, J. (1991) J . Biol. Chem. 266, 31013 106.
9040
Determinants of High-Affinity DNA Binding by the Glucocorticoid Receptor: Evaluation of Receptor Domains outside the DNA-Binding Domain? Karin Dahlman-Wright,' Anthony P. H. Wright, and Jan-Ake Gustafsson Centre for Biotechnology and Department of Medical Nutrition, Karolinska Institute, NO W M , Huddinge University Hospital, S- I41 57 Huddinge. Sweden Received March 2, 1992; Revised Manuscript Received June 18, 1992
In this study, we have investigated the influence of regions outside the DNA-binding domain of the human glucocorticoid receptor on high-affinity DNA binding. We find that the DNA-binding domain shows a 10-fold lower affinity for a palindromic DNA-binding site than the intact receptor. The N-terminal part of the receptor protein does not influence its DNA-binding affinity, while the C-terminal steroid-binding domain increases the DNA-binding affinity of the receptor molecule. It has previously been shown that both the intact glucocorticoid receptor and the glucocorticoid receptor DNA-binding domain bind to a palindromic glucocorticoid response element on DNA as dimers. It is likely that differences in DNA-binding affinity observed result from protein-protein interactions outside the DNA-binding domain betwen receptor monomers, as has been shown for the estrogen receptor. We have previously identified a segment involved in protein-protein interactions between DNA-binding domains of glucocorticoid receptors. This, in combination with results presented in this study, suggests that there are at least two sites of contact between receptor monomers bound to DNA. We suggest that the interaction between the DNA-binding domains may act primarily to restrict DNA binding to binding sites with appropriate half-site spacing and that additional stability of the receptor dimer is provided by the interactions between the steroid-binding domains. ABSTRACT:
The glucocorticoid receptor (GR)' belongs to a family of nuclear receptors which activate transcription in response to binding of their respective ligands. Transcriptional activation is accomplished by DNA-bound receptors. DNA-bindingsites for the GR are called glucocorticoid response elements (GREs) and are characterized by palindromic half-sites which are separated by a three amino acid spacer of variable sequence [forreviews,seeGustafssonetal. (1987),Evans (1988),Beato (1989), and OMalley (1990)l. The palindromic nature of their response elements suggests interaction with receptor dimers, and this has been demonstrated for the GR (Tsai et al., 1988; Wrange et al., 1989; Chalepakis et al., 1990), the estrogen receptor (ER) (Kumar & Chambon, 1988; Fawell et al., 1990), and the thyroid hormone receptor (TR) (Brent et al., 1991; Selmi & Samuels, 199l;Lazaret al., 1991). TheGR(Wrangeet al., 1989), the ER (Kumar & Chambon, 1988; Fawell et al., 1990), and the progesterone receptor (PR) (Rodriguez et al., 1990; DeMarzoet al., 1991)alsoshow somedimerizationin theabsence of DNA. The ER is the best characterized steroid receptor with respect to the segments involved in receptor dimerization. Using ER expressed in yeast cells, it was shown that two segments of the ER are involved in dimerization: one in the steroid-binding domain and one in the DNA-binding domain (Kumar & Chambon, 1988). However, the major dimer?This work was supported by grants From the Swedish Medical Research Council (2819) and the Swedish Natural Science Research Council (K-KU 9756-301). Salary support was provided by the Swedish Medical Research Council to K.D.-W. (student fellowship). * To whom correspondence should be addressed. Telephone: +46 8 6089158. Telefax: +46 8 7745538. Abbreviations: GR, glucocorticoidreceptor;ER, estrogen receptor; TR, thyroid hormone receptor; PR, progesterone receptor;DBD, DNAbinding domain; GRE, glucocorticoid response element; ERE, estrogen response element; DTT, dithiothreitol; TAT, tyrosine aminotransferase.
0006-2960/92/043 1-9040$03.00/0
ization activity was shown to reside in the steroid-binding domain (Kumar & Chambon, 1988; Fawell et al., 1990). A more detailed characterization has revealed the importance of amino acids at the C-terminal end of the steroid-binding domain for ER dimerization (Fawell et al., 1990). The corresponding segment has also been implicated in heterodimerization between retinoic acid and thyroid hormone receptors (Glass et al., 1989). For the PR, mutagenesis studies have indicated a dimerization activity within the ligand-binding domain (Guiochon-Mantel et al., 1989). Finally, it has been suggested that the N-terminal domain of the GR mediates protein-protein interactions between molecules constituting receptor dimers (Eriksson & Wrange, 1990). Two molecules of the GR DNA-binding domain (DBD) bind to a GRE cooperativelydue to protein-protein interactions between bound DBDs (Dahlman-Wright et al., 1990). Mutation of a five amino acid segment close to the second zinc ion has been shown to abolish this activity (Dahlman-Wright et al., 1991). Transient transfection studies using mutated receptor derivatives have shown that this five amino acid segment functions to restrict the transactivation capacity of the glucocorticoid receptor to response elements where the distance between the half-sites is three base pairs (Umesono & Evans, 1989). Thus, protein-protein interactions between DBDs within GR dimers are involved in the specificity of binding site selection. This study was initiated to determine whether the interactions between GR DBDs, in addition to their role in binding site selection, are the sole determinants for GR dimerization. For the ER, it has been shown that receptor dimerization is required for high-affinity DNA binding (Kumar & Chambon, 1988; Fawell et al., 1990). In this study, we have measured the DNA-binding affinity of various GR derivatives to evaluate the role of DBD in relation to other receptor domains in receptor dimerization. 0 1992 American Chemical Society
Biochemistry, Vol. 31, No. 37, 1992 9041
Glucocorticoid Receptor DNA Binding
EXPERIMENTAL PROCEDURES Plasmid Constructions. The yeast expression vector, pKV50, used for expression of GR derivatives has been described (Wright et al., 1990). Using this vector, proteins are expressed from the PGK promoter, the upstream activation sequences of which have been replaced with those from the GAL1,10 promoter. In this expression system, expression is repressed during growth on glucose-containing media and induced on galactose-containing media. The plasmids pKVNX, pKVXE, pKVAX, and pKVNXAT1 have been described previously (Wright & Gustafsson, 1991)and express the following receptor derivatives, respectively: GR (intact receptor, residues 1-777); GRDBD (residues 37&500); GRNtermA (residues 415-777); and GR71A (residues 77-262 deleted). The plasmid pKVNE encodes the GRCtermA protein (residues 1-500) followed by Asp-His-Gly at the Cterminus. Plasmids were transformed into the Saccharomycescerevisiae strain W303-1A (MATa,ade2-1, canl-100, his3-1 I ,15, leu2-3,112, trpl-1, ura3-1) according to Beggs (1978). Growth of Yeast Strains and Preparation of Extracts. Yeast strains were grown to late log phase in SD medium (Sherman et al., 1986) without leucine at 30 OC, harvested by centrifugation, and washed in the same medium without added carbon source. The cells were harvested and resuspended in SGE medium (0.67% yeast nitrogen base without amino acids, 3% glycerol, and 1% ethanol) without leucine, at a density of A600 = 0.5. After 24-h growth at 30 OC, during which glucose repression was relieved, the cells were diluted to a density of A600 = 0.5 in SGE medium without leucine, and galactose was added to a final concentration of 2% to induce expression of the recombinant proteins. Cells were harvested after 8-h growth at 30 "C. Yeast extracts were prepared essentially as described (Wright et al., 1990) except that the extraction buffer was 200 mM Tris-HCI, pH 8.0,400 mM (NH&S04,10 mM MgC12,l mM EDTA, 10%glycerol, 2 mM DTT, and 1 mM PMSF. Extracts were dialyzed against 10 mM Hepes, pH 8.0,10% glycerol, 2.5 mM MgC12,50 mM KCI, 0.1 mM EDTA, and 1 mM DTT. Gel Mobility Shift Assays. For DNA-binding assays, 5 pL of extract was incubated in 10 mM Hepes, pH 8.0, 10% glycerol, 2.5 mM MgC12,50 mM KC1,O. 1 mM EDTA, 1 mM DTT, 1 pg of poly(d1-dC)*poly(dI-dC), and 100 ng of a 41 base pair palindromic estrogen response element (ERE1) described previously (Zilliacus et al., 1991) with various labeled DNA fragments in a total volume of 10 pL. DNA concentrations were determined spectrophotometrically, assuming A260 = 1 correspondsto 20 pg/mL oligonucleotide. The GRE in Figure 1 is derived from the tyrosine aminotransferase (TAT) gene (Jantzen et al., 1987) and is 41 base pairs in length. Oligonucleotideswith inserted or deleted base pairs were correspondingly longer and shorter (Dahlman-Wright et al., 1990). The palindromic GRE studied in Figures 2-5 is a 41 base pair palindromic GRE (GRE1) described previously (Zilliacus et al., 1991). In Figures 1 and 2B, 0.2 ng of labeled oligonucleotidewas added to each incubation. In Figures 3-5, the amount of oligonucleotide added is indicated in the figures. High concentrations of DNA were achieved by diluting the labeled oligonucleotidewith nonlabeled oligonucleotide. Protein-DNA complexes were separated from protein-free DNA by nondenaturingelectrophoresis (Fried & Crothers, 1981) in 5% polyacrylamide gels. The gels were run at 4 OC in 50 mM Tris-boric acid, pH 8.0, 1 mM EDTA, and 0.1 % Triton X- 100 at a constant voltage of 200 V. Following electrophoresis, gels were dried and au-
GRE
A
h
GREM4
GREM5
GREM6
GREM12
h
h
h
h
am--
---
C-, a
I GREM4
(-1GAE)
GREM5
(-[GAGA=)
GREMG
(K~GGACC
FIGURE1: GR dimerization correlates with high-affinity DNA binding. (A) Binding of intact GR to the TAT GRE or variants of this GRE was assayed by band shift analysis. The variant GREs contain 2 (GREM4), 4 (GREMS), or 13 (GREM6) base pairs between the half-sites or have an altered relative orientation of the half-sites (GREM 12). F and C show the positions of the free probe and the GR-DNA complex, respectively. Extracts are diluted 2fold serially from left to right. (B) Sequence of the variant GREs studied in (A). toradiographed. Data were quantified by scintillation counting of bands excised from gels. In some cases, the bands were also quantified by densitometryof autoradiographswhich gave very similar results. The amount of DNA bound was calculated from the specific activity of the probe. Scatchard analysis was used to calculate the dissociation constants for the different proteins. The values measured were reproducible, with individual values varying from the mean of two or more experiments by about 5%, except for the GRNtermA protein where thevariation was 18% from the mean of two experiments. The R values show the reliability of the regression lines used in the Scatchard plots. The Scatchard plots for the GRDBD and GRCtermA proteins are nonlinear at low DNA concentrations, due to cooperativeDNA binding, and, consequently, these points were excluded from the analysis.
RESULTS G R Dimerization Is Required f o r High-Affinity DNA Binding. To evaluate the role of GR dimerization for highaffinity DNA binding, binding of intact GR expressed in yeast cells to a GRE from the TAT gene and variants of this GRE, where the distance between the two half-sites or their relative orientation is changed, was studied using a gel retardation assay (Figure IA). The sequences of the TAT GRE and the variant GREs are depicted in Figure 1B. Binding of the GR, expressed in yeast cells, resulted in the complex marked C in Figure 1A which is specific to the GR since extracts expressing
9042
Biochemistry, Vol. 31, No. 37, 1992
A
DBD
11 r
GR
-_1_1--
.
R
Dahlman-Wright et al.
SBD
1 - y'--
GRDBD GRCtermA
-
I -
-
GRNtermA
B
i
2
3
-"-----T
4
5
o 1.5
e
+ 4:
E 1.0 \ 'D c
;0.5
n
9 0 -.nn
1 .o
0.5
0.0
DNA bound (nM)
. . - - -. . - - - .. . . .
-.
15.0.
;c
A
10.0 .
A
A
5.0 .
: A --..
- * - . .
5
10
. - - - . ' - - . .- -.-
15
20
25
30
DNA (ng)
FIGURE2: Sequence-specific DNA binding by GR derivatives expressedin yeast. (A) GR derivativesexpressed in theyeast extracts utilized in this study. The steroid-binding domain (SBD), the DNAbinding domain (DBD), and the transactivation domains (71 and 72) are indicated. (B) Complexes formed by GR derivatives expressed in yeast using a palindromic GRE (AGAACATGATGTTCT) in a gel mobility shift assay. Extracts contained the expressed proteins as follows: GR (lane l), GRDBD (lane 2), GRCtermA (lane 3), GRNtermA (lanel),and GR71A (lane 5 ) . Fand C show thepositions of the free probe and the protein-DNA complexes,respectively. Note that the lanes shown should not be compared quantitatively since they are taken from DNA titration experiments and contain different amounts of protein and/or DNA.
smaller GR derivatives do not give this complex (Wright & Gustafsson, 1991). Changing the distance between the halfsites constituting a GRE, or changingtheir relative orientation, changes the relative position of receptor molecules on the DNA, potentially interfering with interactions between receptor monomers. Figure 1A shows that the affinity of GR for DNAbinding sites where the distance between the half-sites or their relative orientation has been changed is much reduced. This is consistent with the binding behavior previously shown for the GR DBD (Dahlman-Wright et al., 1990) and supports theview that a dimerized receptor structure, that is compatible with specific interaction with both half-sites of the GRE, is required for high-affinity DNA binding of the GR contained in crude yeast extracts. A similar correlation between halfsite spacing and DNA-binding affinity has previously been demonstrated for GR purified from rat liver (Chalepakis et al., 1990). DNA-BindingAfJinitiesof the Intact GR and the GR DBD. To evaluate the role of the protein-protein contact within the DBD for receptor dimerization and high-affinity DNA binding, the DNA-binding affinities of the intact GR and the GR DBD were compared. The GR and GRDBD proteins expressed in yeast are depicted in Figure 2A. Western blot analysis showed that these proteins, as well as the other GR derivatives used in the study, appeared to be intact in the
3.0
- -
. . . .- . . . . . . . .
:"\ 2.51
D
Kd=5.7 nM
1 .o
0.050
5
10
15
20
DNA bound (nM)
FIGURE 3: DNA-binding affinity of the intact GR and the GR DBD. Scatchard analysis of gel shift assays with yeast extracts containing the GR (B) and the GRDBD (D) proteins. The DNA fragment was a palindromicGRE with sequenceAGAACATGATGTTCT. Shown are the saturation curves for GR and GRDBD (A and C, respectively) and the Scatchard plots for GR and GRDBD (B and D, respectively). The R values for the regression analysis of the Scatchard data were 0.94 and 0.98, respectively.
extracts (data not shown). The DNA-binding activity of the expressed proteins was analyzed using a gel retardation assay. As can be seen in Figure 2B, incubation of extracts expressing GR (lane 1) or GRDBD (lane 2) with a palindromic GRE results in formation of a protein-DNA complex. The migration of the complexes varies according to the size of the receptor derivative in the extracts, and therefore they reflect specific interactions with the GRE. The palindromic GRE used for the affinity determinations was chosen because it shows a lower level of nonspecific binding compared to the TAT GRE (compare Figure 2B and Figure 1A). The DNA-binding affinitiesof the GR and GRDBD proteins were compared by Scatchard analysis of gel retardation experiments using the palindromic GRE (Figure 3). The DNA-binding affinity of the intact GR is about 10-fold higher than that of the isolated DBD, Kd = 0.5 nM versus Kd = 5.7 nM, respectively. This shows that sequencesoutside the DBD are required for high-affinity DNA binding by the GR. These sequences probably mediate protein-protein contacts between receptor monomers that stabilize the protein-DNA complex.
Biochemistry, Vol. 31, No. 37, 1992 9043
Glucocorticoid Receptor DNA Binding
5
: v
5.0 '
O
.
O
F
l
n
n
z
0.3 0.2
2 0.1 0.0 0
5
10
15
20
25
A
0.0 0.0
30
0.5
DNA (no)
5
2.0
Kd-0.95 nM
1.0
E
0.0 0
a i0 0.0 n
0.5 0
1.5
1.0 DNA ( n d
5
10
15
FIGURE 4: DNA-binding affinity of the C-terminally deleted GR. Scatchard analysis of gel shift assays with a yeast extract containing GRCtermA (B). The DNA fragment was a palindromic GRE with sequence AGAACATGATGTTCT. Shown are the saturationcurve (A) and the Scatchard plot (B). The R value for the regression analysis of the Scatchard data was 0.99.
-?
v
I
A
1 .o
Table I: DNA-Binding Affinity of Intact and Truncated OR Proteins GR derivative Kd (nM) GR derivative Kd (nM) GR 0.5 GRNtermA 0.95 GRDBD 5.7 GRT~A 1.4 GRCtermA 5.5
0.0 0.0
1.0
A , 2.0
,
,
3.0
4.0
j
5.0
DNA (no)
ID
-
3.0
Having established the importance of sequences outside the DBD for high-affinity DNA binding, we investigated whether the N-terminal and/or the C-terminal part of the receptor protein contributed to the difference in affinity between the intact GR and the GR DBD. For the ER, the C-terminal steroid-binding part of the receptor protein has been shown to contribute to receptor dimerization and highaffinity DNA binding (Kumar & Chambon, 1988; Fawell et al., 1990). However, for the GR, it has been suggested that protein-protein interactions involving the N-terminal part of the receptor play a role in the interaction of the dimeric receptor with DNA (Eriksson & Wrange, 1990). Role of the G R N- Terminus in High-Affinity DNA Binding. The N-terminal part of the GR (GRCtermA, Figure 2A) was expressed in yeast. Incubation of the extracts with a labeled GRE fragment led to the formation of a protein-DNA complex (Figure 2B, lane 3). This complex represents a specific DNAprotein complex since it was not seen in extracts expressing other GR derivatives (Figure 2B). Quantification of gel retardation experimentsfollowed by Scatchard analysis shows that the DNA-binding affinity of the GRCtermA protein is very similar to that of the GRDBD protein, Kd = 5.5 nM versus Kd = 5.7 nM, respectively (Figures 3 and 4). This shows that sequences N-terminal of the GR DBD do not contain an independent activity that increases the DNAbinding affinity of the receptor protein. Role of the G R C-Terminus in High-Affinity DNA Binding. The lack of dimerization activity in the N-terminal domain suggests that the C-terminal ligand-binding domain of the GR contributes to receptor dimerization, as has been shown for the ER (Kumar & Chambon, 1988; Fawell et al., 1990). To test this directly, receptor derivatives from which the whole N-terminus (GRNtermA) or part of the N-terminus (GR71A) were deleted, shown in Figure 2A, were expressed and the
IC
'IAA, 3.0
1 .o
0.5 DNA bound (nM)
0.0
DNA bound (nM)
0.0 0.0
1.0
2.0
3.0
4.0
5.0
DNA bound (nM)
FIGURE5: DNA-binding affinity of N-terminally deleted GR derivatives. Scatchard analysis of gel shift assays with yeast extracts containing the proteins GRNtermA (B) and GRrlA (D). The DNA fragment was a palindromic GRE with sequence AGAACATGATGTTCT. Shown are the saturation curves for GRNtermA and G R T ~ A(A and C, respectively) and the Scatchard plots for GRNtermA and GRTlA (B and D, respectively). The R values for the regression analysis of the Scatchard data were 0.96 and 0.97, respectively.
extracts used to produce the specific complexesshown in Figure 2B (lanes 4 and 5 ) . These proteins bound to the GRE with dissociation constants of 0.95 and 1.4 nM, respectively, indicatingthat the C-terminus does contribute to high-affinity DNA binding by the GR. The two proteins gave very similar results in spite of the low level of active protein in the GRNtermA extracts. This probably results from aggregation of the GRNtermA protein when it is overexpressed (Wright et al., 1990) and causes an increase in nonspecific binding to the GRE (Figure 2B), thus making quantitation more difficult. DISCUSSION
In Figure 1A, we show that high-affinity DNA binding by the GR is practically abolished when the distance between, or the relative orientation of, the two half-sites constituting the GRE is altered. This probably reflects an influence on the ability of the receptor to form DNA-bound dimers as has previously been suggested (Chalepakis et al., 1990). We have used a gel retardation assay, followed by Scatchard analysis, to evaluate the role of different parts of the GR
9044
Biochemistry, Vol. 31, No. 37, 1992
for maximal DNA-binding affinity, as summarized in Table I. The results show that regions outside of the DBD are required for maximal DNA-binding affinity by the GR (Figure 3). It is likely that this increased DNA-binding affinity is the result of protein-protein interactions between receptor monomers bound to the GRE. This has previously been demonstrated for the ER (Kumar & Chambon, 1988; Fawell et al., 1990). However, we cannot exclude that regions outside the DBD mediate direct contactswith DNAor that they indirectly influence DNA-binding affinity by, for example, causing structural changes within the DBD. The N-terminal region of the GR has previously been implicated in receptor dimerization (Eriksson & Wrange, 1990). However, this domain does not appear to increase the DNA-binding affinity of the receptor and thus would not be expected to mediate contacts between receptor monomers (Figure 4), at least not in the absence of the steroid-binding domain. Addition of the C-terminal steroid-binding domain increases the DNA-binding affinity of the DBD (Figure 5). This suggests that this part of the receptor molecule contributes protein-protein interactions as has previously been shown for the ER (Kumar & Chambon, 1988; Fawell et al., 1990). The DNA-binding affinity of the GRNtermA and GR71A proteins is not quite as high as that of the intact GR. Thus, the N-terminus seems to be required to achieve maximal DNAbinding affinity. This could reflect an auxiliary role of the N-terminus that enhances the activity in the C-terminus and would be consistent with previous observations that the N-terminus does play a role in receptor dimerization (Eriksson & Wrange, 1990). Alternatively, since the contribution of the N-terminus is very small, it might be due to a general artifact associated with theuse of abnormal truncated proteins. At present, we cannot distinguish between these possibilities. However, we can conclude that the C-terminus of the GR makes a major contribution to the receptor’s DNA-binding affinity and that the N-terminus cannot contribute independently, although it might perform an auxiliary function in the intact GR. In vitro mutagenesis has revealed the importance of amino acids at the C-terminal end of the steroid-binding domain for ER dimerization. A striking feature of this sequence is a heptad repeat of hydrophobic residues with additional hydrophobic residues at intermediate positions (Fawell et al., 1990). The segment including these amino acids, particularly the repeating pattern of hydrophobic residues, is conserved within the nuclear receptor family, and it will be interesting to see if the same amino acids are involved in high-affinity DNA binding by the GR.
Dahlman-Wright et al.
REFERENCES Beato, M. (1989) Cell 56, 335-344. Beggs, J. (1978) Nature 275, 104-108. Brent, G. A., Williams, G. R., Harney, J. W., Forman, B. M., Samuels, H. H., Moore, D. D., & Larsen P. R. (1991) Mol. Endocrinol. 5, 542-548. Chalepakis,G., Schauer, M., Cao, X.,& Beato, M. (1990) DNA Cell Biol. 9, 355-368. Dahlman-Wright, K., Siltala-Roos, H., Carlstedt-Duke, J., & Gustafsson, J.-A. (1990) J. Biol. Chem. 265, 14030-14035. Dahlman-Wright,K., Wright, A., Gustafsson,J.-A., & CarlstedtDuke, J. (1991) J. Biol. Chem. 266, 3107-3112. DeMarzo, A. M., Beck, C. A., Onate, S.A., & Edwards, D. P. (1991) Proc. Natl. Acad. Sci. U.S.A. 88, 72-76. Eriksson, P., & Wrange, 0. (1990) J. Biol. Chem. 265, 35353542.
Evans, R. M. (1988) Science 244, 889-895. Fawell, S. E., Lees, J. A., White, R., & Parker, M. (1990) Cell 60, 953-962.
Fried, M., & Crothers, D. M. (198 1) Nucleic Acids Res. 8,65056525.
Glass, C. K., Lipkin, S. M., Devary, 0.V., & Rosenfeld, M. G. (1989) Cell 59, 697-708. Guiochon-Mantel, A., Loosfelt, H., Lescop, P., Sar, S., Atger, M., Perrot-Applanat,M., & Milgrom,E. (1989) Ce1157,11471 154.
Gustafsson, J.-A., Carlstedt-Duke, J., Poellinger, L., Okret, S., Wikstrbm,A X . , Brbnnegiird, M., Gillner, M., Dong, Y.,Fuxe, K., Cintra, A,, Hlrfstrand, A., & Agnati, L. (1987) Endocr. Rev. 8, 185-234. Hollenberg, S. M., Weinberger, C., Ong, E. S., Cerelli, G., Oro, A., Lebo, R., Thompson, E. B., Rosenfeld, M. G., & Evans (1985) Nature 318, 635-641. Jantzen, H.-M., Strlhle, U., Glass, B., Stewart, F., Schmid, W., Boshart, M., Miksicek, R., & Schiitz, G. (1987) Cell 49,2938.
Kumar, V., & Chambon, P . (1988) Cell 55, 145-156. Lazar, M. A., Berrodin, T. J., & Harding, H. P. (1991) Mol. Cell. Biol. 11, 5005-5015. Ohara-Nemoto, Y.,Strbmstedt, P.-E., Dahlman-Wright, K., Nemoto, T., Gustafsson, J.-A., & Carlstedt-Duke, J. (1990) J. Steroid Biochem. Mol. Biol. 37, 481-490. O’Malley, B. W. (1990) Mol. Endocrinol. 4, 363-369. Rodriguez, R., Weigel, N., O’Malley, B., & Schrader, W. T. (1990) Mol. Endocrinol. 4, 1782-1790. Selmi, S., & Samuels, H. H. (1991) J. Biol. Chem. 266,1158911593.
Sherman, F., Fink, G. R., & Hicks, J. B. (1986) Laboratory Course Manualfor Methods in Yeast Genetics, 2nd ed., Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. Tsai,S., Carlstedt-Duke,J., Weigel, N., Dahlman, K., Gustafsson, J.-A., Tsai, M.-J., & O’Malley, B. (1988) Cell 55, 361-369. Umesono,-K., & Evans, R. M. (1989) Cell 57, 1139-1146. Wrange, O., Eriksson, P., & Perlmann, T. (1989) J. Biol. Chem. 255, 5253-5259.
ACKNOWLEDGMENT We thank the members of the Steroid Receptor Unit at the Centre for Biotechnology for useful discussions and advice, Helmi Siltala for excellent technical assistance, and Delta Biotechnology Limited for the pKV5O yeast expressionvector.
Wright, A. P. H., & Gustafsson, J.-A. (1991) Proc. Natl. Acad. Sci. U.S.A. 88, 2883-2887. Wright, A. P. H., Carlstedt-Duke, J., & Gustafsson,J.-A. (1 990) J. Biol. Chem. 265, 14763-14769. Zilliacus, J., Dahlman-Wright, K., Wright, A,, Gustafsson, J.A., & Carlstedt-Duke, J. (1991) J . Biol. Chem. 266, 31013 106.
